Diabetic retinopathy is associated with hyperglycemia, and there is convincing evidence that oxidative stress (the overproduction of reactive oxygen species (ROS)) measured indirectly in patients, is related to the severity of diabetic complications. Also, reducing such stress by various means, including drugs and reducing hyperglycemia, decreases the rate of development of diabetic retinopathy (DR). It is therefore supposed that oxidative stress causes DR, and the injuries caused by ROS in retinas of diabetic animals (and in human preparations). But the cells first affected in small animal models of DR are found throughout the inner retina, and not specifically associated with small blood vessels until later in the development of retinopathy. We raise the questions A: why in human disease the small retinal blood vessels are so selectively affected B: what are the processes that induce oxidative damage in the retina. There are difficulties in interpretation of experimental results, because there is no metric which relates the degree of damage to the level of ROS, either in clinical or animal experiments and the relative sensitivity of the different methods employed to demonstrate oxidative damage in experiments on tissues is unquantified. It is also important to note that in addition to oxidative stress, hyperglycemia induces several changes including leucostasis, vasoconstriction and a pro-inflammatory state that also causes hypoxia in the retina. The earliest retinal pathology and the earliest biochemical changes appear to begin within 1 week of the time when the animals become diabetic and are provoked by hyperglycemia. These changes include alterations to the appearance of microglia, the formation of Advanced Glycation Endproducts (AGEs), the overproduction of Vascular Endothelial Growth Factor (VEGF) and its mRNA and consequent leakage of capillary endothelial cells. These early pro-inflammatory changes can directly cause hypoxia in the retina and not necessarily via ROS. Experiments on isolated cells indicate that retinal capillaries are less susceptible to hyperglycemia than other retinal cells, but in vivo are selectively damaged, possibly via paracrine changes. This suggests a new concept: although the changes in blood vessels may be a consequence of gradual and cumulative development of oxidative stress, the preceding paracrine and other changes that cause the development of oxidative stress are highly significant to the understanding and treatment of DR. The clinical importance is that about the time that oxidative stress becomes easily demonstrable, the progress of DR is already irreversible. A number of methods of treatment of DR depend upon the relief of retinal hypoxia. If oxidative stress is considered 'the' determinant of DR, explanation of such findings solely in terms of oxidative stress would require additional hypotheses.